Week4.Rmd

---
title: "Week4"
output: html_document
---

```{r}
moviematrix <- read.csv2("./week4_movierow.csv")
movienames <- moviematrix[, 1]
movierow <- (moviematrix[, -1])
rownames(movierow) <- movienames
usercor <- cor(movierow, movierow, use = "pairwise.complete.obs")

user_avg <- colMeans(movierow, na.rm=T)
movierow_normalized <- t( t(movierow) - user_avg)
#3867 and 89
for(x in c("X3712", "X3867", "X89")){
  cat("Nearest neighbors ")
  cat(x)
  cat(":\r\n")
  cor_order <- order(usercor[x, ], decreasing=T)[-1][1:5]
  cor_vals <- usercor[x, cor_order]
  print(cor_vals)

  cat("Top movies ")  
  cat(x)
  cat(":\r\n")
  {
    weighted_ratings <- t( t(movierow[, cor_order]) * cor_vals)
    applied_weights <- t( t(!is.na(weighted_ratings)) * cor_vals)
    predicted_ratings <- rowSums(weighted_ratings, na.rm = T) / rowSums(applied_weights)
    top_movies <- sort(predicted_ratings, decreasing = T)[1:5]
    print(top_movies)
  }
  
  cat("Top normalized movies ")
  cat(x)
  cat(":\r\n")
  {
    weighted_ratings <- t( t(movierow_normalized[, cor_order]) * cor_vals)
    applied_weights <- t( t(!is.na(weighted_ratings)) * cor_vals)
    predicted_ratings <- rowSums(weighted_ratings, na.rm = T) / rowSums(applied_weights)
    predicted_ratings <- predicted_ratings + user_avg[x]
    top_movies <- sort(predicted_ratings, decreasing = T)[1:5]
    print(top_movies)
  }
}
```

```{r}
```